Is cognitive behaviour therapy applicable to individuals diagnosed with bipolar depression or suboptimal mood stabilizer treatment: a secondary analysis of a large pragmatic effectiveness trial

Abstract Background Efficacy trials of medications and/or psychological interventions for bipolar disorders (BD) aim to recruit homogenous samples of patients who are euthymic and such populations show high levels of adherence to the treatments offered. This study describes a secondary analysis of a large-scale multi-centre pragmatic effectiveness randomized controlled trial (RCT) of cognitive behaviour therapy plus treatment as usual (CBT) or treatment as usual alone (TAU) and explores outcomes in individuals who were: (i) recruited in depressive episodes, or (ii) receiving suboptimal doses of or no mood stabilizers (MS). Methods Data were extract on two separate subsamples (out of 253 RCT participants). Sample 1 comprised 67 individuals in a depressive episode (CBT: 34; TAU: 33); Sample 2 comprised 39 individuals receiving suboptimal MS treatment (CBT: 19; TAU: 20). Survival analyses (adjusted for confounding variables) were used to explore recovery in Sample 1 and relapse in Sample 2. Results In Sample 1 (individuals with depression), Cox proportional hazards regression model revealed that the median time to recovery was significantly shorter in the CBT group (10 weeks; 95% confidence intervals (CI) 8, 17) compared to the TAU group (17 weeks; 95% CI 9, 30) [Adjusted Hazard Ratio (HR) 1.89; 95% CI 1.04, 3.4; p < 0.035]. In Sample 2 (suboptimal MS), the median time to any relapse was significantly longer in the CBT group compared to the TAU group (~ 35 versus ~ 20 weeks; Adjusted HR 2.01; 95% CI 1.01, 3.96; p < 0.05) with the difference in survival time to first depressive relapse also reaching statistical significance (X2 = 14.23, df 6, p 0.027). Conclusions Adjunctive use of CBT appears to have benefits for individuals diagnosed with BD who are highly representative of the patients seen in routine clinical practice, but often excluded from efficacy RCTs. However, as this is a secondary analysis of 42% of the original RCT sample, it is important to replicate these findings in independent larger scale studies specifically designed for purpose.